v0.58.2

This release brings critical fixes for a crash in Apricot, and for an issue with PartitionInitialize in the ODC client plugin. It also fixes an issue with missing SOR timestamps in the Bookkeeping client.

• [apricot] Notify about verbose setting
• [apricot] Don't crash because of nil getStringMap
• [core] Apricot host-detector inventory cache
• [core] Avoid direct Apricot defaults/vars calls from ODC client
• [core] Include envId in pre-deploy cleanup failure message
• [core] Bookkeeping plugin corrects missing SOR timestamps in case they are not set correctly.
• [core] Pass the env's defaults+vars from Consul to ObjectStack context

v0.58.1

This release brings a single critical fix for environment deployment.

• [core] Do not fail env deployment on cleanup issues

v0.58.0

This release includes bug fixes and integration improvements.

• Bug fixes:

  • [core] Fix crash in pre-deployment cleanup error handling
  • [core] Pre-deployment cleanup failure is now a WARNING
  • [executor] Manually handle buffering in task stdout logging
  • [OCTRL-649][core] Correcting MIXED/ERROR logic after PR review.
  • [OCTRL-649][core] Aggregator role should have the correct state between MIXED and ERROR, if it has multiple children roles in those states.

• Integration:

  • [OCTRL-704][core] template helpers for assigning JIT workflows per groups of FLPs
  • [OCTRL-709][core] pass flp id list to ccdb grp object
  • [OCTRL-744][core] The BK plugin functions GetData and GetEnvironmentsData return pending runs to be stopped.
  • [OCTRL-744][core] GetData and GetEnvironmentsData functions return O2 and TRG pending stop info available in the BK plugin.

• Miscellaneous:

  • [core] HostnameWithin -> SuffixInRange

v0.57.0

This release includes bug fixes and improvements to AliECS core functionality.

• Apricot:

  • [apricot] Add method logging to service

• Deployment fixes:

  • [core] Impose a 10s timeout on each ssh cleanup call
  • [core] Make task reuse optional, and disabled by default
  • [core] LeafWalk should also walk in include roles
  • [core] workflow.Walk needs to recurse without LeafWalk

• Logging:

  • [core] 'task inactive exception' should be an error log
  • [executor] truncate task command in logs when it will be cut

v0.56.1

The present release includes bug fixes and integration improvements.

• [core] Fix crash in TRG plugin
• [core] Remove pdp_epn_shm_recreate variable from ODC client
• [core] Keep GEN_TOPO_WIPE_CACHE outside the MD5 cache hash
• [core] Fix appending GEN_TOPO_WIPE_CACHE to PDP script call

v0.56.0

This release includes a refactored client for the O² Bookkeeping service, migrated from OpenAPI to a gRPC-based interface, a new mechanism for exposing integrated services data to AliECS clients, support for non-critical data flow tasks, deployment performance improvements, and miscellaneous improvements and fixes.

• Non-critical data flow tasks:

  • [OCTRL-649] First modifications for allowing non-critical tasks to go into ERROR without the system failing.
  • [OCTRL-649] Second iteration of implementation of non-critical tasks.
  • [OCTRL-649] The workflow state is updated correctly when a non-critical is in ERROR.

• Integrated services data:

  • [coconut] Integrated services data in coconut env show
  • [coconut] Suppress warning
  • [core] Add integrated services data to EnvironmentInfo in proto
  • [core] Refactor GetData and implement GetEnvData in all plugins
  • [core] Include integrated services data in all EnvironmentInfo payloads

• Bookkeeping:

  • [core] Core Bookkeeping plugin is now prepared to be gRPC based in synch. with Bookkeeping.
  • [core] The AliECS Bookkeeping plugin is migrated to gRPC, matching the Bookkeeping implementation.
  • [core] Detectors are correctly parsed from string to Detector enum type to match Bookkeeping proto.
  • [core] The AliECS core Bookkeeping plugin is migrated to gRPC, matching Bookkeeping.
  • [core] BK run logs are created even if run failed to start, and run end times are set if missing when a run is stopped.
  • [core] The Makefile points to the correct BK branch.
  • [core] The run number is attached to AliECS Bookkeeping plugin messages.

• Deployment performance:

  • [core] Bring back optional concurrency in workflow processing
  • [core] Concurrent WFT expansion defaults to false for now

• Miscellaneous:

  • [core] Amend EPN script call generator according to OCTRL-736
  • [executor] do not try to recover a task in error, ask to exit instead
  • [OCTRL-718] Leading/trailing whitespaces are trimmed from Consul KV imported variables if the flag trimSpaceInVarsFromConsulKV is true.
  • [OCTRL-718][apricot] Consul vars are correctly trimmed if they contain whitespaces.

v0.55.1

This patch release includes a fix for pre-deployment cleanup, as well as improved RPC server logging.

• [core] Disable strict host checking in pre-deploy cleanup ssh call
• [core] Enable and expand RPCserver method logging and time tracking

v0.55.0

The present release brings important improvements to log output, as well as a concurrency refactor in the AliECS core with the goal of fixing deployment issues and improving performance.

• Scheduler concurrency:

  • [core] Protect crit section in scheduler and make more concurrent
  • [core] Improve scheduler debug output
  • [core] Correct scope of local offer

• Miscellaneous:

  • [build] Bump gRPC dependency
  • [core] Fix crash in incoming event handling
  • [core] log task name when launching, do not log unknown fields
  • [core] logging failed task should be an error, add the task name

v0.54.3

This patch release further fixes the CTP readout enabled scenario.

• [core] CCDB client: include TRG in GRP object if ctp_readout_enabled
• [misc] Improve documentation of RunType enums

v0.54.2

The present patch release includes an API fix for environment-related calls.

• [core] Ensure consolidated defaults/vars/userVars are sent back